The role of source and timing on social media while
using a denial crisis-response strategy.
Master’s Thesis
Graduate School of Communication, University of Amsterdam
Corporate Communication
Author:
Anton Bekenkamp (10266208)
Supervisor:
dr. G.L.A. (Toni) van der Meer
Abstract
Social media play an increasingly important role in crisis communication. Past literature has shown the potential of employees as a source in crisis communication. Moreover, the timing of disclosing crisis-related information is proven to be an important factor. This study builds on previous research by investigating the role of source and timing while using a denial crisis-response strategy. A 2 (source: employee vs. organization) x 2 (timing: stealing thunder vs. thunder) experiment was conducted amongst 164 participants. The results show that
employees have a more positive effect on reputation compared to the organization. However, the crisis source did not impact secondary crisis communication. When investigating timing, a stealing thunder timing strategy did not impact reputation and secondary crisis
communication. Lastly, the effect of source on reputation and secondary crisis
communication is found to be mediated by crisis responsibility. This study provides relevant insights for communication practitioners and functions as a starting point for future research in the field of crisis communication.
Introduction
Social media play an increasingly important role in crisis communication. The public relies on social media to obtain and share crisis-related information (Lachlan, Spence, Lin, Najarian & Greco, 2016), while organization’s use these new technologies in their strategic
communication processes (Macnamara & Zerfass, 2012).
Crises can occur in any organization and are characterized as unexpected events that impact the organization and its stakeholders, damaging the firm’s reputation (Coombs, 2006). Hence, crisis management is critical to limit harm to the organization’s stakeholders or society in general (Coombs & Holladay, 2002; Coombs, 2007). Crisis communication is, for example, crucial to ensure public safety and welfare, avoid lawsuits and prevent financial loss (Coombs & Holladay, 2002). Crisis-response strategies can help companies to minimize reputational damage.
As outlined by situational crisis communication theory (SCCT), organizations have a variety of crisis-response strategies to choose from (Coombs, 2006, 2007). When actual or potential harm is done to the organization’s stakeholders, a victim-centered approach is recommended, addressing concerns about the stakeholders’ safety (Coombs, 2014). In practice however, denial is the most used strategy by organizations (Kim, Avery & Lariscy, 2009). Denial can be an effective strategy to minimize reputational damage in situations where the organization holds no crisis responsibility (Coombs, 2007). Nevertheless,
organizations are frequently involved in the crisis to some extent. If the organization is shown to have any connection to a crisis event while using a denial strategy, reputational damage is intensified (Coombs, 2014).
Although it may be counterintuitive for practitioners to admit the organization’s involvement or to disclose negative crisis-related information, an aggressive communication approach is desirable over a passive approach (Moran & Gregory, 2014; Coombs, 2014). The
timing of the crisis communication may be a crucial factor in this approach, to endeavor the desired effects of the crisis response strategy.
Past literature has addressed the concept of stealing thunder as a beneficial timing strategy in crisis communication (e.g. Arpan & Pompper, 2003; Coombs, 2015).
Organizations can use this strategy with the objective to ‘break the news’ about a crisis event before other parties do, resulting in less reputational damage (Arpan & Roskok-Ewoldsen, 2005; Claeys & Cauberghe, 2012). The organization’s position may be weakened if third parties, such as the news media, initially disclose the crisis, allowing their frame to become dominant. Yet, research towards the effects of timing while using a denial strategy is largely absent. Therefore, the current research builds on the previous findings by examining the effects of timing when the organization denies any involvement in the crisis.
Moreover, the recent body of literature has addressed the role of social media in crisis communication in more depth (e.g. Procopio & Procopio, 2007; Schultz, Utz & Göritz, 2011; Lin, Spence, Sellnow & Lachlan, 2016). Social media has become an important vessel in crisis communication for disseminating information from the organization to the affected publics (Freberg, 2012). In line, other research has proven that crisis communication through social networks result in a higher reputation and less secondary crisis communication
compared to online newspapers (Utz, Schultz & Glocka, 2013).
The advent of social media allows many actors other than the organization to actively participate in the crisis conversation by sharing ‘opinions, insights, experiences, and
perspectives with others’ (Marken, 2007). Recent studies highlighted the importance of employees as third-party senders of crisis communication (Van Zoonen, van der Meer & Verhoeven, 2014; Van Zoonen & van der Meer, 2015), functioning as valuable brand ambassadors of the organization (Cravens & Oliver, 2006; Dreher, 2014). Consequently, employees could be considered as valuable assets in crisis communication as they can
influence other stakeholders. Past crisis communication research (i.e. Van Zoonen & van der Meer, 2015) concerning different sources on social media has not addressed the role of timing. Also, little is known about the potential of employees in crisis communication while using a denial strategy.
This research will contribute to established and recent crisis communication theories focusing on social media. Further this study wants to contribute to SCCT (Coombs, 2006) by further developing the role of denial crisis-response strategies when the organization’s involvement in a crisis is still unknown or debatable. The results may provide practitioners with relevant implications on how to prevent reputational damage when a crisis hits an
organization. The following research question is formulated: In times of crisis, can source and timing help the organization to minimize reputational damage when using a denial crisis-response strategy on social media?
Theoretical Framework
Impact of Crises on Organization’s
Organizational crises are characterized as unexpected and harmful events affecting the organization on an operational or reputational level (Coombs & Holladay, 2002; Coombs, 2006). Events such as industrial and natural disasters can disrupt the daily operation of an organization, and form a threat to public safety or the well-being of stakeholders (Sohn & Lariscy, 2013). On the other hand, causalities such as mismanagement, human errors or technological failures damage the firm’s reputation. Crises cause stakeholders to re-evaluate their impressions about an organization and frequently result in a financial loss
(Zyglidopoulos & Phillips, 1999). Whenever a crisis hits an organization, the primary goal is to protect the organization and its stakeholders from threats inflicted by the crisis (Coombs, 2007). The organization’s response can limit or repair reputational damage done by a crisis (Coombs, 2002, 2006). Hence, selecting the appropriate response strategy can help managers
to take control of the crisis and minimize harm to the organization (Coombs, 2002). In past literature, scholars have focused on several key variables potentially affected by
organizational crises.
First, as mentioned before, crises impact reputation as one of the main assets for organizations. Organizational reputation can be defined as the “stakeholders perceptions about an organization's ability to create value relative to competitors” (Rindova, Williamson, Petkova & Sever, 2005). Reputation is linked with perceived value, customer satisfaction and loyalty (Booker & Serenko, 2007). Moreover, organizational reputation is developed through interactions between the public and the organization or its employees (Van Zoonen & van der Meer, 2015). Hence, understanding how reputation is affected by crisis communication is essential for effective crisis management.
Secondly, several studies argue that secondary crisis communication affects how stakeholders evaluate the organization in post-crisis communication. Schultz, Utz and Göritz (2011) define secondary crisis communication as the “intentions to tell friends about the crisis, to share the received information with others and to leave comments”. Secondary crisis communication is comparable with word-of-mouth and can be either beneficial for the
organization (positive word-of-mouth) or cause potential damage (negative word-of-mouth) (Coombs & Holladay, 2008; Laczniak, DeCarlo & Ramaswami, 2001). Minimizing negative secondary crisis communication could therefore be an effective strategy to protect the
organization from reputational harm. Firms can also try to control secondary crisis
communication, for example by encouraging their audiences to spread official press releases aimed at rebuilding the organization’s reputation (Lin, Spence, Sellnow & Lachlan, 2016). Moreover, social media enable stakeholders to share crisis related information with ease, stimulating secondary crisis communication. Hence, managing secondary crisis
Situational Crisis Communication Theory (SCCT)
Communication in times of crisis can manage the stakeholder’s expectations and provide crisis managers with options to limit reputational damage, negative secondary crisis communication and improve crisis response acceptance. SCCT grants a framework to
minimize reputational harm in post-crisis communication by proposing several crisis-response strategies that can be used to achieve this goal (Coombs, 2007). According to Coombs (1995), crisis response strategies are “what an organization says and does after a crisis hits”. Coombs and Holladay (2002) identify three crisis clusters, defining the crisis responsibility of the firm. The degree of crisis responsibility affects the reputational threat, along with crisis history and prior relationship reputation (Coombs, 2007). In the victim cluster, the organization is a victim of the crisis event and holds no responsibility. In the accidental cluster, the crisis was caused by the organization unintentionally. In turn, the intentional cluster reflects a situation where the organization has a high responsibility for the crisis event.
Moreover, SCCT defines three groups of crisis-response strategies. Denial strategies focus on rejecting any connection between the organization and crisis event. Second, diminish strategies focus on assuring stakeholders the organization did not intend to do any harm or claiming the crisis was not in control of the organization. Last, rebuild strategies aim to restore the organization’s’ reputation by offering material or symbolic forms of aid to victims (Coombs, 2007). Coombs and Holladay (2008) argue that organizational reputation is
influenced by the crisis-response strategy.
Denial as Crisis-response Strategy
The current research will focus on denial crisis-response strategies, being the most frequently used by organization’s (Kim et al., 2009). Organization’s tend to use this strategies to avoid
legal actions and financial claims (Tyler, 1997). Although denial is the most frequently used strategy, it was also evaluated as the least effective strategy because it poses a risk when perceived as inappropriate or fails to satisfy the stakeholders’ needs (Grebe, 2013). According to Coombs (2007), organizations should prioritize protecting their stakeholders from harm done by a crisis. While diminish and rebuild strategies are accommodative towards the victims, denial strategies are defensive and organization-oriented. This might explain why a denial strategy is not effective in most cases, since the main focus of this strategy is to disclaim the organization’s guilt.
However, a denial strategy can be appropriate in given situations. For example, research by Van der Meer (2014) found that a denial strategy can be effective when the organization is not responsible for the crisis. In this case, the denial strategy resulted in frame adoption and led to the acceptance of the organization’s response. Based on these findings, denial strategies are only considered effective if the crisis fits the victim cluster. Denial strategies are for example appropriate in misinformation crises, when the organization is victimized by untrue information and holds no responsibility (Coombs, 2014). Uncontrollable factors (i.e. natural disasters or malevolence) can also justify the use of denial (Coombs, 2007). However, in the initial phase of a crisis it is often unclear who holds responsibility and thus which cluster the crisis fits. Crisis responsibility is partly determined in communication when assigning meaning to crisis events (Claeys & Cauberghe, 2014).
Crisis Communication on Social Media
Stakeholders can turn to a variety of sources to make sense of a crisis (Palen, Vieweg, Liu & Hughes, 2009). In crisis communication, social media can help to rebuild the stakeholders’ trust (Derani & Naidu, 2016) and diminish the impact of a crisis (Wendling, Radisch & Jacobzone, 2013; Yates & Paquette, 2011). On the other hand, stakeholders often use social
networks to gather relevant information about crises while evaluating the associated risks and responsibilities (Valentini & Romenti, 2011). Furthermore, social media services can be utilized to control negative word-of-mouth (Tucker & Melewar, 2005) and prevent damage to the brand’s reputation caused by secondary crisis communication (Coombs & Holladay, 2007). Twitter is particularly used to share crisis-related information with the affected community (Smith, 2010; Mendoza, Poblete & Castillo, 2010).
Employees as Source of Crisis Communication
Because social media allow anyone to be a content creator (Flanagin & Metzger, 2000), new sources of crisis-communication have emerged.Third parties such as employees are often forgotten as a potential source of crisis communication, despite their crucial role in protecting the organization in times of crisis (Van Zoonen & van der Meer, 2015). Traditionally,
employees relied on internal communication systems and the long-established mass media to voice their positive or negative experiences with the organization (Miles & Mangold, 2014). Presently, social media enable employees to communicate their satisfaction or dissatisfaction to a broad audience without effort. Miles and Mangold (2014) refer to this phenomenon as employee voice, defined as “an employee’s attempt to use either organizationally sanctioned or unsanctioned media or methods for the purpose of articulating organizational experiences and issues or influencing the organization, its members, or other stakeholders”. While not all enterprises have embraced employees as valuable communication assets, other organization’s provided their employees with social media guidelines, encouraging employees to use social media to the company’s advantage (Barker, 2008). More importantly, employees are
perceived as authentic representatives of their organization and therefore act as powerful online influencers (Dreher, 2014).
In crisis communication, employees are both receivers and senders of information, participating in social networks inside and outside of the organization (Frandsen & Johansen, 2011). Employees are different compared to other stakeholders, as they hold a close
relationship with the organization (Frandsen & Johansen, 2011; Johansen, Aggerholm & Frandsen, 2012). Employees see the organization as part of their identity (Wiesenfeld, Raghuram & Garud, 2000), resulting in an obligatory feeling of protecting the organization from reputational damage (Frandsen & Johansen, 2011). According to Van Zoonen, van der Meer and Verhoeven (2014), employees can use social media to build relationships with stakeholders and positively frame the organization online. In line, employees are well aware of the possible negative effects of work-related tweets and are thus likely to tweet neutral and factual information about the organization (Van Zoonen, Verhoeven & Vliegenthart, 2016). It is expected that employees will not harm the organization in times of crisis when
communicating on social media. Instead, the literature suggests employees are valuable assets in protecting the organization in times of crisis, preventing reputational harm. The credibility of employees will lead to acceptance of the denial response-strategy. Thus, it is expected secondary crisis communication will decrease. This leads to the following hypotheses:
H1a: In case of a denial strategy, an employee as crisis source leads to a higher reputation compared to the organization as a source.
H1b: In case of a denial strategy, an employee as crisis source leads to decreased secondary crisis communication compared to the organization as a source.
Timing in Crisis Communication
The timing of releasing crisis related information can be crucial (Arpan & Pompper, 2003; Arpan & Roskos-Ewoldsen, 2005; Coombs, 2007). Timing refers to the moment when information acknowledging the existence of a crisis is released (Coombs, 2015). Particularly in the early stage of a crisis, such as the prodromal period (Fink, 1986), choosing the right timing strategy can prevent further organizational damage. News media often put pressure on organizations to provide information quickly (Veil & Ojeda, 2010) due to the high news value of crisis situations (Kleinnijenhuis, Schultz, Utz & Oegema, 2013). Moreover, crises can be interpreted differently amongst stakeholders. After a crisis hits, sense-making processes unfold rapidly resulting in a variety of frames surrounding the crisis event (Van der Meer, Verhoeven, Beentjes, & Vliegenthart, 2013). Timely communication is, therefore, necessary for the organization to become part of the frame building process and facilitate crisis
understanding. Existing literature about timing and crisis communication often refers to the concept of ‘stealing thunder’ as a timing strategy. The concept was first addressed in literature in the field of law, referring to stealing thunder as a “persuasion tactic in which an individual reveals potential incriminating evidence first, for the purpose of reducing its negative impact evaluative audience” (Dolnik, Case & Williams, 2003).
In crisis communication, an organization is ‘stealing thunder’ when it is the first to report about an crisis incident before other sources do (Arpan & Pompper, 2003; Coombs, 2007). The opposite of stealing thunder is referred to as ‘thunder’ (Williams, Bourgeois & Croyle, 1993; Dolnik et al., 2003), meaning others had discovered the crisis before the organization itself acknowledged it. This is, for example, the case when journalists release crisis-related information before the organization does. A stealing thunder strategy positively impacts the organization’s’ credibility and diminishes reputational damage (Claeys &
by organizations to steal thunder and enable crisis managers to quickly report a crisis without relying on traditional news media (Coombs, 2014). Moreover, a timely response using
stealing thunder gives the organization more control of its response strategy, preventing news agencies from spreading speculation about a crisis event (Ihlen, 2002). Last, accommodative response strategies (i.e. rebuild strategies) are deemed less necessary if the organization steals thunder (Claeys & Cauberghe, 2012). This suggests that a defensive strategy (i.e. denial) is effective when stealing thunder. In a situation where crisis responsibility is still unclear, refuting any connection between the organization and the crisis using a defensive response and stealing thunder timing strategy is expected to positively influence reputation. Moreover, it is expected that stealing thunder leads to less secondary crisis communication because stakeholders are more likely to accept the organization’s denial. Hence, the following hypotheses are formulated:
H2a: In case of a denial strategy, a stealing thunder timing strategy leads to a higher reputation than a thunder timing strategy
H2b: In case of a denial strategy, a stealing thunder timing strategy leads to less secondary crisis communication than a thunder timing strategy
Source Credibility
In crisis communication, the credibility of the source is one of the predictors when evaluating reputation (Van Zoonen and van der Meer, 2015). Credibility refers to the audience’s
confidence and acceptance of the source and its message (Hovland, Irving & Harold, 1968). The current research will focus on the credibility of the source, or in other words, the
credibility of the communicator (Van Zoonen & van der Meer, 2015; Greer, 2003; Sundar & Nass, 2001).
With regards to crisis communication on social media, stakeholders evaluate the credibility of the sender of crisis-related information (i.e. the employee or organization). Expertise and trustworthiness of the crisis source are important factors used by audiences in this evaluation process (Ibelma & Powell, 2001; Hovland et al., 1982). Moreover, the credibility of the source influences how audiences form their opinion about the received information(Beach, Mitchell, Deaton & Prothero, 1978). Hence, crisis managers should be aware of the influence source credibility has on the organization's stakeholders. Credibility is especially important when a denial crisis-response strategy is used. If the crisis source is perceived as credible, the relevant audiences might accept the claim that the organization holds no responsibility. Therefore, it is expected that the effect of crisis source and timing on organizational reputation and secondary crisis communication is mediated by the credibility of the source. The following hypotheses are formulated:
H3a: The main effect of crisis source on reputation is mediated by the credibility of the source.
H3b: The main effects of crisis source on secondary crisis communication is mediated by the credibility of the source.
H3c: The main effect of timing on reputation is mediated by the credibility of the source.
H3d: The main effects of timing on secondary crisis communication is mediated by the credibility of the source.
Crisis responsibility
Stakeholders can have different opinions or thoughts about the extent to which the organization is responsible for a crisis. Coombs and Holladay (2005) define crisis
responsibility as “how much people believe the organization is responsible for the crisis”. Crisis responsibility is not about the organization’s actual guilt, but about the audience’s perception of the firm’s responsibility (Benoit, 1997). Attribution theory explains that people have the need to search for causes of negative and unexpected events (Wiener, 1986).
Consequently, stakeholders determine crisis responsibility based on the attributions made about the cause of the crisis (Coombs, 2007).
To minimize crisis responsibility, organization’s should take notice of the dominant frames amongst their stakeholders and to what extent the firm is held responsible. It is
expected that the effect of source and timing on organizational reputation and secondary crisis communication is mediated by the perceived crisis responsibility.
H4a: The main effect of crisis source on reputation is mediated by crisis responsibility.
H4b: The main effect of crisis source on secondary crisis communication is mediated by crisis responsibility.
H4c: The main effect of timing on reputation is mediated by crisis responsibility.
H4d: The main effect of timing on secondary crisis communication is mediated by crisis responsibility.
Crisis response acceptance
Third, the effectiveness of the crisis response strategy may predict if the organization’s’ response is perceived as appropriate by the publics. Acceptance of the crisis communication strategy may limit reputational harm. Jin (2010) defines crisis response acceptance as “how publics accept organization’s crisis strategy”. In line, Coombs and Holladay (2008) refer to this concept as account acceptance, meaning “how respondents feel about the crisis response offered by the organization”. According to SCCT, crisis response acceptance is closely linked with response strategies (Coombs, 2007). Whether stakeholders accept the
crisis-response from the organization, influences to what extent the crisis-response strategy affects reputation (Coombs & Holladay, 2008). Therefore, it is expected crisis response acceptance functions as an important mediator, leading to the following hypotheses:
H5a: The main effect of crisis source on reputation is mediated by crisis response acceptance.
H5b: The main effect of crisis source on secondary crisis communication is mediated by response acceptance.
H5c: The main effect of timing on reputation is mediated by crisis response acceptance.
H5d: The main effect of timing on secondary crisis communication is mediated by crisis response acceptance.
Method
Participants
An experiment was conducted amongst 192 participants. Participants have been recruited trough social media and the researchers own network and were incentivized for their
participation by offering the chance to win a cinema voucher. It was ensured that the subjects were familiar with Twitter. Respondents who did not meet this requirement were excluded. Eventually, the final sample consisted of 164 subjects (39.6% male, 60.4% female) with a mean age of 26 years (SD = 11.34). Of all participants, 69.5% (N = 114) followed or completed higher education (university).
Procedure and Design
To answer the research question, a 2 (crisis source: employee vs. organization) x 2 (timing strategy: stealing thunder vs. thunder) between-subjects experimental design was used to examine the main effects on organizational reputation and secondary crisis communication. In addition, source credibility, crisis responsibility, and crisis response acceptance were included as mediating variables to assess possible indirect effects.
The experiment was conducted using an online questionnaire. First, participants were informed about the nature of the research and their consent for participation was asked. Since the experimental material involves a Twitter profile, participants were asked if they were familiar with Twitter. Only subjects acquainted with Twitter were allowed to continue. Next, participants were familiarized with a fictitious company called ‘Best Coffee’. A short briefing explained that Best Coffee is a well-known chain of coffee stores operating in The
Netherlands, serving coffee and other hot and cold beverages, sandwiches and snacks. Further, the crisis was briefly introduced by explaining how a customer claimed to got sick after visiting one of Best Coffee’s stores, but that it remains unclear if Best Coffee caused the
illness. Thus, crisis responsibility was open for the respondents to determine. Next, the participants were randomly assigned to one of the four experimental conditions. Four conditions were created, varying in the source of the crisis communication and the used timing strategy (employee/thunder, employee/stealing thunder, organization/thunder, organization/stealing thunder). Subsequently, organizational reputation, secondary crisis communication, crisis response acceptance, source credibility and crisis responsibility were measured. Finally, a manipulation check was conducted, and demographic data were asked. On the last page, the participants were thanked for their participation, and it was explained that the organization, Twitter profile, and crisis were fictional and only created for the current research.
Experimental Materials
For the current experiment, four different Twitter profile pages were fabricated, related to a fictitious company called ‘Best Coffee’. The crisis case was held constant in all conditions and reported a customer got ill due to a bacterial infection after visiting a Best Coffee store. The source was manipulated by randomly assigning the respondents to either a Twitter profile owned by Best Coffee or an employee. To maintain a high internal validity, several elements of the Twitter profile were adjusted to make a clear distinction between the organization and the employee.
The source was manipulated by altering the profile picture, header image and the biography of the user’s profile. For example, the employee’s profile featured a personal profile picture while the brand’s profile included their logo. Further, the profile’s statistics such as the number of tweets and followers was changed. As is often the case with
professional accounts, the companies profile was verified by Twitter (indicated by a blue badge next to the username) and included a link to Best Coffee’s website.
Timing was manipulated by altering the message of the first Tweet. The thunder condition included a retweet (sharing another account's Tweet) of a news article published by the NOS, a well-known Dutch public service news broadcaster. The official NOS Twitter-account (@NOS) tweeted the following message: “Best Coffee customer claimed to got sick from dangerous bacteria. Read the full article here: nos.nl/29382”. This retweet was
accompanied with a response from either the organization or the employee. The organization responded: “According to the @NOS, a customer got sick due to a bacterial infection after visiting Best Coffee.”, while the employees Twitter message stated: “I just read this article from @NOS, allegedly a customer got sick after visiting my store”. In turn, the stealing thunder condition did not include a retweet from the news agency, but the initial report about the incident was done by the employee or organization. In the employee source condition, the Tweet stated: “While working at Best Coffee, a customer called us and claimed to got sick after a visit to my store. Something with bacteria.”. On the contrary, the organization broke the news by Tweeting: “Our customer service just got an alarming phone call. One of our customers claimed to got sick by a bacterial infection.”.
To make sure the manipulation of the source and timing was perceived correctly, a briefing before the manipulation also disclosed the owner of the Twitter profile (employee or organization) and who initially reported the crisis incident (employee, organization or news media).
Moreover, a denial strategy was used by both the employee and the organization. Three more tweets simulated this strategy by denying the involvement of Best Coffee with the incident, claiming the stores were recently checked and approved by the authority of food safety and that this was the first time a customer got sick.
The questionnaire and experimental material were translated into Dutch (Appendix D), as the research was conducted in The Netherlands.
Dependent Measurements
Organizational reputation was measured using six item 7-point Likert-scale adapted from the research of Coombs and Holladay (2002) and Fombrun, Gardberg and Sever (2000).
Respondents were asked to what extend they agreed (1=fully disagree, 7=fully agree) with items such as: “I trust Best Coffee”, “Best Coffee offers high quality products” and “Best Coffee is a responsible company”. A new scale was created by calculating the means of the six items of organizational reputation (Cronbach’s α = .90, M = 4.32, SD = 1.07).
Secondary crisis communication was measured using a three item 7-point Likert-scale adapted from the research of Schultz et al. (2011). Respondents were asked to indicate the likeliness (1=very unlikely, 7=very likely) of: sharing the tweets with others (retweeting), tell others about the tweets and placing a reaction on the tweets. A new scale was created by calculating the means of the these items (Cronbach’s α = .75, M = 2.15, SD = 1.21).
Source credibility was measured using a five item 7-point Likert-scale based on the research of Metzger, Flanagin, and Zwarun (2003) and Grewal, Gotlieb and Marmorstein (1994). Respondents were asked to what extend they agreed (1=fully disagree, 7=fully agree) with items such as: “The sender of the tweets is credible”, “The sender of the tweets can be trusted” and “The sender of the tweets is an expert”. A new scale was created by calculating the means of the five items (Cronbach’s α = .81, M = 3.79, SD = 1.11).
Crisis responsibility was measured by six items using a 7-point Likert-scale adapted from the research of Brown and Ki (2013). Respondents were asked to what extend they agreed (1=fully disagree, 7=fully agree) with items such as: “The cause of the crisis was intentionally done by someone (or something) within the organization.”, “The crisis was preventable by the organization.” and “The organization should be blamed for the crisis.”. An initial reliability analysis suggested an unreliable scale, Cronbach’s α = .52. Therefore, two
items negatively affecting the scale were deleted. A new scale was created with the remaining four items, by calculating the means of the crisis responsibility items (Cronbach’s α = .73, M = 3.21, SD = 1.04).
Crisis response acceptance was measured by five items using a 7-point Likert-scale adapted from the research of Jin (2010). Respondents were asked to what extend they agrede (1=fully disagree, 7=fully agree) with items such as: “The reaction of the Twitter user was appropriate”, “The Twitter user acted correctly” and “The reaction of the Twitter user was sincere”. A new scale was created by calculating the means of the five items (Cronbach’s α = .93, M = 3.66, SD = 1.35).
Pretest
A pretest was conducted amongst 25 test participants to examine if the source and timing manipulations were perceived as intended. Using an online questionnaire, respondents were randomly assigned to one of the four conditions. First, to check the manipulation of the source, respondents were asked to identify the owner of the Twitter profile as either an employee or the organization. Moreover, to verify the manipulation of timing, respondents were asked whether the employee, organization or news media was the first to report about the crisis incident. A second question asked if the owner of the Twitter profile was the one to announce the news about the crisis. Respondents could also select “don’t know” for each question.
Appendix A summarizes the manipulation of source and timing. A chi-square test showed significant differences between the groups for the manipulation of source X2 (6) = 17.34, p = 0.008 and the manipulation of timing X2 (9) = 32.67, p = 0.000.
Furthermore, the participants were asked to rate the authenticity of the experimental material on a 7-point Likert scale. Results show the material was perceived as sufficient
authentic (M = 5,56, SD = 0.87). Additionally, 84% of the respondents indicated to be familiar with Twitter. Hence, the manipulation material was not altered in the proceeded experiment.
Data Analysis
Several statistical analysis were conducted to answer the defined hypothesis. For hypothesis1 and 2, a multivariate analysis of covariance (MANCOVA) was conducted. This analysis was found most suitable as the dependent and mediating variables were measured on interval level. Moreover, an MANCOVA it allows measurements for multiple dependent variables and the inclusion of covariates. Based on the conditions, dummy variables were created for source (0 = organization, 1 = employee) and timing (0 = thunder, 1 = stealing thunder). Further, to test the effects of the mediators (hypothesis 3, 4 and 5), different regression models were constructed using the PROCESS mediation analysis (Hayes, 2013). PROCESS is a regression-based approach to test multiple regressions within one model. Moreover, the PROCESS analysis uses bootstrapping to ensure reliable and accurate results for the
analytical models. Hence, the PROCESS analysis is found to be most suitable for the current research as it allows testing of direct and indirect effects of multiple independent and
mediating variables.
Results
Randomization check
To check if the respondents are equally distributed amongst the four experimental conditions, a one-way analysis of variance (ANOVA) was conducted with the four conditional groups as the independent variable and age as the dependent variable. None of the four conditions showed significant differences for age F (3, 159) = 1.74, p = .16. Further, a chi-square test
show insignificant differences for educational level between the conditions X2 (12) = 15.55, p = .213. Further, another chi-square test showed no significant differences for gender between the conditions X2 (3) = 3.68, p = 0.298. Last, it was checked if participants possessing a Twitter account were divided equally amongst the groups. Again, no significant differences between the conditions were found X2 (3) = 0.46, p = .927. Based on this analysis, it is assumed that there are no differences between the conditions considering the variables age, educational level, gender and Twitter account possession.
The participants were roughly equally divided amongst the conditions. Condition A (employee/thunder) counted 38 participants (23,2%), condition B (employee/stealing thunder) counted 43 participants (26,2%), condition C (organization/thunder) counted 42 participants (25,6%) and last 41 participants (25%) were assigned to condition D (organization/stealing thunder).
Manipulation check
In the current research, the source and timing of the Twitter messages were altered. To check if the manipulation of the source was successful, a control question asked whether the Twitter profile was owned by an employee or the organization. A chi-square test showed significant differences between the groups X2 (6) = 116.71, p < .001, meaning the manipulation of the source was successful. To check the manipulation of the timing, respondents had to indicate who communicated about the crisis incident first: the news media, an employee or the organization. A chi-square test checking the manipulation of timing showed significant differences between the groups X2 (9) = 148.33, p < .001, suggesting successful manipulation. Appendix B summarizes the results.
Covariates
A Pearson correlation matrix was constructed to control for the variables gender, age,
education and Twitter account in correspondence with the dependent variables organizational reputation and secondary crisis communication. No significant correlations were found between the control and dependent variables. Therefore, gender, age, education and Twitter account were not included in further analysis. In addition, the mediator’s source credibility, crisis responsibility and crisis response acceptance were included in the matrix. Significant correlations were found. Hence, these variables were added as covariates in further analysis. Appendix C shows an overview of the correlations and the significant effects.
Hypothesis testing
The hypothesis formulated in the current research are tested below.
Hypothesis 1a and 1b. First, testing the effect of crisis source, the MANCOVA found
a marginally significant effect for crisis source on organizational reputation, F (1, 157) = 3.05, p = .083. Subjects who were exposed to the employee source condition scored marginally significant higher on organizational reputation (M = 4.39, SD = 1.04) compared to subjects exposed to the organization source condition (M = 4.26, SD = 1.10). However, no significant effect was found for the effect of crisis source on secondary crisis communication, F (1, 157) = 2.36, p = .127. Subjects exposed to the employee source condition did not score
significantly higher on secondary crisis communication (M = 2.23, SD = 1.31) compared to subjects exposed to the organization source condition (M = 2.07, SD = 1.10). The results suggest that using an employee as source of crisis communication results in a higher organizational reputation while using a denial strategy then when the organization is the source, accepting hypothesis 1a. However, these effects do not hold up for secondary crisis
Hypothesis 2a and 2b. Second, testing the effects of timing, the MANCOVA showed
no significant effect was found for timing on organizational reputation, F (1, 157) = .24, p = .624. Subjects exposed to the thunder timing strategy did not score significantly higher on organizational reputation (M = 4.28, SD = 1.18) compared to subjects exposed to the stealing thunder timing strategy (M = 4.36, SD = 0.96). Furthermore, no significant effect was found for timing on secondary crisis communication F (1, 157) = .46, p = .499. Subjects exposed to the thunder timing strategy did not score significantly higher on secondary crisis
communication (M = 2.18, SD = 1.34) compared to those exposed to the stealing thunder timing strategy (M = 2.12, SD = 1.21). The results suggest that using an stealing thunder timing strategy does not result in a higher organizational reputation or less secondary crisis communication while using a denial strategy. Hence, hypothesis 2a and 2b are rejected.
Interaction source and timing. Furthermore, the ANCOVA revealed no significant
interaction effects of source and timing on reputation, F (1, 157) = .12, p = .707 and secondary crisis communication, F (1, 157) = 1.49, p = .224. The means of the four conditions on the dependent variables are summarized in Table 1. Although no significant effects were found, the mean differences seem to be in the desired direction: when an
employee steals thunder, organizational reputation is perceived higher. Moreover, the stealing thunder conditions both result in less secondary crisis communication.
Table 1. Means of interaction effect source and timing on dependent variables
Organizational reputation Secondary crisis communication
Condition M SD M SD
Employee / Thunder 4.36 1.23 2.39 1.62 Employee / Stealing Thunder 4.42 .84 2.09 .94 Organization / Thunder 4.30 1.07 2.14 1.20 Organization / Stealing Thunder 4.21 1.14 1.99 1.00
Last, Levene’s test assumed equal variances in the population for the variable organizational reputation, F (3, 160) = 1.59, p = .194 and crisis response acceptance F (3, 160) = 0.91, p = .436. It should be noted that the assumption of equal variances in the population has been violated for the variable secondary crisis communication, F (3, 160) = 2.81, p = .041.
Mediation analysis
To test hypothesis 3, 4, and 5, five regression models were constructed using the PROCESS mediation analysis (Hayes, 2013). The first three models tested the effects of the independent variables (source and timing) on the mediating variables (source credibility, crisis
responsibility and crisis response acceptance). The fourth model examined the effects of the mediating variables and independent variables on organizational reputation while the fifth model examined the effects on secondary crisis communication. When constructing each of the models, either source or timing was added as covariate. Later the covariate and
independent variable were switched to accurately measure the direct effects of source and timing on reputation and secondary crisis communication. The findings of the models are summarized in Table 2. Finally, Sobel tests were conducted to check for significant mediation effects.
Effects on source credibility, crisis responsibility and response acceptance.
Initial analysis revealed that the regression model of crisis source and timing with the mediator source credibility as the outcome was not significant (Model 1). Source and timing do not predict source credibility. Further, the regression model of source and timing with crisis responsibility as the outcome was significant (Model 2), but the strength of the prediction is weak. Only crisis source is found to be the significant predictor. Third, crisis
source and timing are not significant predictors for crisis response acceptance (Model 3). The model is not found to be significant.
Effects on reputation.
The overall model (Model 4) with the effects of the independent and mediating variables on organizational reputation was significant and has a strong prediction. It is suggested that source credibility, crisis responsibility, and crisis response acceptance are significant predictors for organizational reputation. Crisis source shows a marginally significant effect, while no effect was found for timing. For all these effects, other variables are assumed to be held constant.
Hypothesis 3a, 4a, and 5a. When looking at the direct effects of source on reputation
while ignoring the mediators and including timing as covariate, the regression was not significant (b* = .13, t (2,161) = 0.78, p = .438). Results of the Sobel test suggests that the relation between crisis source and organizational reputation is marginally significant mediated by crisis responsibility (Sobel Z = 1.80, p = 0.073) but not by source credibility (Sobel Z = -.58, p = 0.564) and response acceptance (Sobel Z = -1.50, p = 0.134). Based on these findings, hypothesis 4a is accepted. However, no significant mediation effect was found for source credibility and response acceptance on reputation, rejecting hypothesis 3a and 5a.
Hypothesis 3c, 4c, and 5c. Moreover, the regression of timing on reputation, ignoring
the mediators and including source as covariate, was not significant (b* = .08, t (2,161) = 0.46, p = .661), suggesting no direct effect of timing on reputation. The Sobel test suggests that the relation between timing and organizational reputation is not significantly mediated by source credibility (Sobel Z = .85, p = 0.398), crisis responsibility (Sobel Z = .68, p = 0.495) and response acceptance (Sobel Z = -.96, p = 0.335). Based on these results, hypothesis 3c, 4c and 5c are rejected.
Effects on Secondary Crisis Communication
The overall model (Model 5) with the effects of the independent and mediating variables on secondary crisis communication was significant but has a weak prediction. The model suggests that source credibility and crisis responsibility are significant predictors for secondary crisis communication. Response acceptance, crisis source, and timing are however not significant. For all these effects, other variables are assumed to be held constant.
Hypothesis 3b, 4b, and 5b. Looking at the direct effects of source on secondary crisis
communication while ignoring the mediators and including timing as covariate, the regression was not significant (b* = .16, t (2,161) = .87, p = .388). Results of the Sobel test suggest that the relation between crisis source and secondary crisis communication is marginally
significant mediated by crisis responsibility (Sobel Z = -1.76, p = 0.079) but not by source credibility (Sobel Z = -.56, p = 0.574) and response acceptance (Sobel Z = .60, p = 0.550). Taken together, these results suggest crisis responsibility is a possible mediator in the relation between source and secondary crisis communication, accepting hypothesis 4b. No significant mediation effect was found for source credibility and response acceptance on secondary crisis communication, rejecting hypothesis 3b and 5b.
Hypothesis 3d, 4d, and 5d. The regression of timing on secondary crisis
communication, ignoring the mediators and including source a covariate, was not significant (b* = .07 t (2,161) = -.36, p = .720), suggesting no direct effect of timing on secondary crisis communication. A Sobel test suggests that the relation between timing and secondary crisis communication is not significantly mediated by source credibility (Sobel Z = .82, p = 0.413), crisis responsibility (Sobel Z = -.68, p = 0.498) and response acceptance (Sobel Z = .47, p = 0.636). Based on these results, hypothesis 3d, 4d and 5d are rejected.
Table 2. Regression models to predict dependent variables.
Note. N = 164 * p < 0.05
** p < 0.10 (marginally significant)
Model b* t p R2 F df
1 Outcome: Source credibility
Overall model .587 .01 .53 2, 161
Crisis source -.10 -.59 .558
Timing .15 .87 .388
2 Outcome: Crisis responsibility
Overall model .048* .04 3.09 2, 161
Crisis source -.38 -2.35 .020*
Timing -.12 -.74 .461
3 Outcome: Response acceptance
Overall model .166 .02 1.82 2, 161 Crisis source -.34 -1.60 .112 Timing -.21 -1.00 .319 4 Outcome: Reputation Overall model .000* .56 40.53 5, 158 Source credibility .41 6.13 .000* Crisis responsibility -.17 -3.01 .003* Response acceptance .28 5.09 .000* Crisis source .20 1.75 .082** Timing .05 .48 .634
5 Outcome: Secondary crisis communication
Overall model .000* .14 5.21 5, 158 Source credibility .41 3.86 .000* Crisis responsibility .25 2.87 .005* Response acceptance -.07 -.76 .448 Crisis source .28 1.53 .129 Timing -.11 -.63 .532
Conclusion and Discussion
The current research experimentally investigated the effects of crisis source and timing on social media while using a denial strategy. The effects on organizational reputation and secondary crisis communication were measured. Moreover, the role of source credibility, crisis responsibility and crisis response acceptance as possible mediators was explored.
First, it was expected that an employee as crisis source leads to a higher organizational reputation than when the organization was the source. The results support this expectation as employees are found to have a more positive effect on reputation than when the organization was the source, when using a denial strategy. These findings are in line with Van Zoonen, van der Meer and Verhoeven (2014), who argued that employees positively influence the
organization on social media. Further, the favorable impact employees have on reputation might be accounted for by their authentic and credible appearance (Dreher, 2014).
Second, it was expected that an employee as crisis source decreased secondary crisis communication compared to the organization as a source, while using a denial strategy. However, this expectation was not supported by the current results. Although Schultz, Utz and Goritz (2010) found that post-crisis communication influences secondary crisis communication, the current results suggest consumers will not engage in more or less secondary crisis communication irrespectively of the source.
Third, it was expected that a stealing thunder timing strategy results in a higher organizational reputation than a thunder timing strategy. However, the current study implies that timing does not impact organizational reputation. The results are in contrary with the findings of Claeys and Cauberghe (2011) who argued that a stealing thunder strategy
positively impacts the organization’s’ credibility, diminishing reputational damage. The fact that Claeys and Cauberghe (2011) focused on a preventable crisis while providing factual
information instead of denying the organization’s involvement, might explain these
differences. Future research on stealing thunder as timing strategy must decide if strategies other than denial yield the same results.
Moreover, it was expected that a stealing thunder strategy leads to less secondary crisis communication. However, secondary crisis communication was not affected by the timing strategy. This is somewhat in line with the research by Arpan and Pompper (2003), who suggested that the severity of the crisis was not impacted by timing. Hence it is possible secondary crisis communication is linked with the gravity of the crisis.
In line, previous literature suggested that stealing thunder results in a higher credibility of the source (Eagly, Wood and Chaiken, 1978; Arpan and Pompper, 2003). As the current results prove otherwise, it could be that this effect does not hold up when using denial strategies.
Lastly, the results show that the effect of crisis source on reputation and secondary crisis communication is mediated by crisis responsibility. When the organization was the source, subjects assigned more responsibility to the organization compared to the employee as source. This could be explained by the fact that employees are seen as powerful online
influencers (Dreher, 2014). Thus, when employees use denial in their response, the results suggest the organization’s responsibility for the crisis decreases, leading to a higher reputation. Further, if the organization is found to be more responsible for the crisis, it is likely secondary crisis communication will increase. However, the effect of timing on reputation and secondary crisis communication was not mediated by crisis response
acceptance. The results further suggest that the effect of source and timing on reputation and secondary crisis communication is also not mediated by source credibility and crisis
In conclusion, when organization’s in times of crisis use denial as their response strategy on social media, employees should be considered as valuable actors in helping the organization to minimize reputational damage. Timing, however, is not found to reduce reputational damage when using a denial. Further research is needed to examine the effects of timing in more detail.
Limitations & practical implications
One limitation of the current research is the use of a fictitious organization and crisis. Future research could examine the effects of source and timing in an actual crisis, as Coombs (2007) suggests that prior relational reputation and crisis history impacts the reputational thread. Moreover, the current experiment was mostly conducted amongst students. Although this group is likely to use and be familiar with social media platforms, it is possible that elderly or less familiar consumers process information on social media differently. Further, Twitter was especially examined in this study. However, it could be that the effects of source and timing while using a denial strategy are different when applied in other online environments (i.e. blogs, Facebook, corporate website).
The current research has several implications for crisis communication and public relation practitioners. First, this research extends the SCCT by demonstrating the importance of employees in crisis communication while using a denial strategy. Crisis managers could use this knowledge to improve their strategic communication processes. For example, as suggested by Barker (2009), offering social media guidelines for employees could prepare them before a crisis hits. These guidelines could help organizations to use employees in crisis communication to their full potential. Furthermore, the present results emphasize the
importance of crisis responsibility. With the emerging social media, many actors can engage in the crisis discourse, framing the responsibility of the crisis. Therefore, companies should
not only be aware their actual responsibility but also if their audiences see the organization as responsible.
References
Arpan, L. M., & Pompper, D. (2003). Stormy weather: testing “stealing thunder” as a crisis communication strategy to improve communication flow between organizations and journalists. Public Relations Review, 29(3), 291-308.
doi:10.1016/s0363-8111(03)00043-2
Arpan, L. M., & Roskos-Ewoldsen, D. R. (2005). Stealing thunder: Analysis of the effects of proactive disclosure of crisis information. Public Relations Review, 31(3), 425-433. doi:10.1016/j.pubrev.2005.05.003
Barker, P. (2008). How social media is transforming employee communications at Sun Microsystems. Global Business and Organizational Excellence, 27(4), 6-14. doi:10.1002/joe.20209
Beach, L. R., Mitchell, T. R., Deaton, M. D., & Prothero, J. (1978). Information relevance, content and source credibility in the revision of opinions. Organizational Behavior and Human Performance, 21(1), 1-16. doi:10.1016/0030-5073(78)90034-x
Benoit, W. L. (1997). Image repair discourse and crisis communication. Public Relations Review, 23(2), 177-186. doi:10.1016/s0363-8111(97)90023-0
Brown, K. A., & Ki, E. (2013). Developing a Valid and Reliable Measure of Organizational Crisis Responsibility. Journalism & Mass Communication Quarterly, 90(2), 363-384. doi:10.1177/1077699013482911
Claeys, A., & Cauberghe, V. (2012). Crisis response and crisis timing strategies, two sides of the same coin. Public Relations Review, 38(1), 83-88.
doi:10.1016/j.pubrev.2011.09.001
Claeys, A., & Cauberghe, V. (2014). What makes crisis response strategies work? The impact of crisis involvement and message framing. Journal of Business Research, 67(2), 182-189. doi:10.1016/j.jbusres.2012.10.005
Coombs, T. W. (2014). State of Crisis Communication: Evidence and the Bleeding Edge. Research Journal of the Institute for Public Relations, 1(1), 1-12. Retrieved from http://www.instituteforpr.org/state-crisis-communication-evidence-bleeding-edge/
Coombs, W. T. (2007). Protecting Organization Reputations During a Crisis: The
Development and Application of Situational Crisis Communication Theory. Corporate Reputation Review, 10(3), 163-176. doi:10.1057/palgrave.crr.1550049
Coombs, W. T. (1995). Choosing the Right Words: The Development of Guidelines for the Selection of the "Appropriate" Crisis-Response Strategies. Management
Communication Quarterly, 8(4), 447-476. doi:10.1177/0893318995008004003 Coombs, W. T., & Holladay, S. J. (2002). Helping Crisis Managers Protect Reputational
Assets: Initial Tests of the Situational Crisis Communication Theory. Management Communication Quarterly, 16(2), 165-186. doi:10.1177/089331802237233
Coombs, W. T., & Holladay, S. J. (2005). An Exploratory Study of Stakeholder Emotions: Affect and Crises. Research on Emotion in Organizations The Effect of Affect in Organizational Settings, 1, 263-280. doi:10.1016/s1746-9791(05)01111-9
Coombs, W. T. (2006). The Protective Powers of Crisis Response Strategies. Journal of Promotion Management, 12(3-4), 241-260. doi:10.1300/j057v12n03_13
Coombs, W. T., & Holladay, S. J. (2007). The negative communication dynamic: Exploring the impact of stakeholder affect on behavioral intentions. Journal of Communication Management, 11(4), 300-312. doi:10.1108/13632540710843913
Coombs, W. T., & Holladay, S. J. (2008). Comparing apology to equivalent crisis response strategies: Clarifying apologys role and value in crisis communication. Public Relations Review, 34(3), 252-257. doi:10.1016/j.pubrev.2008.04.001
Coombs, W. T. (2015). Ongoing crisis communication: planning, managing, and responding. Thousand Oaks, CA: SAGE.
Cravens, K. S., & Oliver, E. G. (2006). Employees: The key link to corporate reputation management. Business Horizons, 49(4), 293-302. doi:10.1016/j.bushor.2005.10.006 Derani, N. E., & Naidu, P. (2016). The Impact of Utilizing Social Media as a Communication
Platform during a Crisis within the Oil Industry. Procedia Economics and Finance, 35, 650-658. doi:10.1016/s2212-5671(16)00080-0
Dolnik, L., Case, T. I., & Williams, K. D. (2003). Stealing thunder as a courtroom tactic revisited: Processes and boundaries. Law and Human Behavior, 27(3), 267-287. doi:10.1023/a:1023431823661
Dreher, S. (2014). Social media and the world of work. Corporate Communications: An International Journal, 19(4), 344-356. doi:10.1108/ccij-10-2013-0087
Eagly, A. H., Wood, W., & Chaiken, S. (1978). Causal inferences about communicators and their effect on opinion change. Journal of Personality and Social Psychology, 36(4), 424-435. doi:10.1037//0022-3514.36.4.424
Fink, S. (1986). Crisis Management - Planning for the Inevitable. New York, NY, AMACON (American Management Association).
Flanagin, A. J., & Metzger, M. J. (2000). Perceptions of Internet Information Credibility. Journalism & Mass Communication Quarterly, 77(3), 515-540.
doi:10.1177/107769900007700304
Fombrun, C. J., Gardberg, N. A., & Sever, J. M. (2000). The Reputation QuotientSM: A multi-stakeholder measure of corporate reputation. Journal of Brand Management, 7(4), 241-255. doi:10.1057/bm.2000.10
Frandsen, F., & Johansen, W. (2011). The study of internal crisis communication: towards an integrative framework. Corporate Communications: An International Journal, 16(4), 347-361. doi:10.1108/13563281111186977
Freberg, K. (2012). Intention to comply with crisis messages communicated via social media. Public Relations Review, 38(3), 416-421. doi:10.1016/j.pubrev.2012.01.008
Grebe, S. K. (2013). Things can get worse: How mismanagement of a crisis response strategy can cause a secondary or double crisis: the example of the AWB corporate scandal. Corporate Communications: An International Journal, 18(1), 70-86.
doi:10.1108/13563281311294137
Greer, J. D. (2003). Evaluating the Credibility of Online Information: A Test of Source and Advertising Influence. Mass Communication and Society, 6(1), 11-28.
doi:10.1207/s15327825mcs0601_3
Grewal, D., Gotlieb, J., & Marmorstein, H. (1994). The Moderating Effects of Message Framing and Source Credibility on the Price-Perceived Risk Relationship. Journal of Consumer Research, 21(1), 145. doi:10.1086/209388
Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis: a regression-based approach. New York: The Guilford Press. Hovland, C. I., Janis, I. L., & Kelley, H. H. (1968). Communication and persuasion:
psychological studies of opinion change. New Haven: Yale University Press. Ibelema, M., & Powell, L. (2001). Cable Television News Viewed as Most Credible.
Ihlen, O. (2002). Defending the Mercedes A-Class: Combining and Changing Crisis-Response Strategies. Journal of Public Relations Research, 14(3), 185-206. doi:10.1207/s1532754xjprr1403_2
Jin, Y. (2010). Making Sense Sensibly in Crisis Communication: How Publics’ Crisis
Appraisals Influence Their Negative Emotions, Coping Strategy Preferences, and Crisis Response Acceptance. Communication Research, 37(4), 522-552.
doi:10.1177/0093650210368256
Johansen, W., Aggerholm, H. K., & Frandsen, F. (2012). Entering new territory: A study of internal crisis management and crisis communication in organizations. Public Relations Review, 38(2), 270-279. doi:10.1016/j.pubrev.2011.11.008
Kim, S., Avery, E. J., & Lariscy, R. W. (2009). Are crisis communicators practicing what we preach?: An evaluation of crisis response strategy analyzed in public relations research from 1991 to 2009. Public Relations Review, 35(4), 446-448.
doi:10.1016/j.pubrev.2009.08.002
Kleinnijenhuis, J., Schultz, F., Utz, S., & Oegema, D. (2015). The Mediating Role of the News in the BP Oil Spill Crisis 2010. Communication Research, 42(3), 408-428. doi:10.1177/0093650213510940
Lachlan, K. A., Spence, P. R., Lin, X., Najarian, K., & Greco, M. D. (2016). Social media and crisis management: CERC, search strategies, and Twitter content. Computers in Human Behavior, 54, 647-652. doi:10.1016/j.chb.2015.05.027
Laczniak, R. N., Decarlo, T. E., & Ramaswami, S. N. (2001). Consumers’ Responses to Negative Word-of-Mouth Communication: An Attribution Theory Perspective. Journal of Consumer Psychology, 11(1), 57-73. doi:10.1207/s15327663jcp1101_5
Lin, X., Spence, P. R., Sellnow, T. L., & Lachlan, K. A. (2016). Crisis communication, learning and responding: Best practices in social media. Computers in Human Behavior, 65, 601-605. doi:10.1016/j.chb.2016.05.080
Macnamara, J., & Zerfass, A. (2012). Social Media Communication in Organizations: The Challenges of Balancing Openness, Strategy, and Management. International Journal of Strategic Communication, 6(4), 287-308. doi:10.1080/1553118x.2012.711402 Marken, A. G. (2007). Social Media... The Hunted Can Become the Hunter. Public Relations
Meer, T. G., Verhoeven, P., Beentjes, H., & Vliegenthart, R. (2014). When frames align: The interplay between PR, news media, and the public in times of crisis. Public Relations Review, 40(5), 751-761. doi:10.1016/j.pubrev.2014.07.008
Mendoza, M., Poblete, B., & Castillo, C. (2010). Twitter under crisis. Proceedings of the First Workshop on Social Media Analytics - SOMA 10. doi:10.1145/1964858.1964869
Metzger, M. J., Flanagin, A. J., & Zwarun, L. (2003). College student Web use, perceptions of information credibility, and verification behavior. Computers & Education, 41(3), 271-290. doi:10.1016/s0360-1315(03)00049-6
Miles, S. J., & Mangold, W. G. (2014). Employee voice: Untapped resource or social media time bomb? Business Horizons, 57(3), 401-411. doi:10.1016/j.bushor.2013.12.011 Moran, R., & Gregory, J. R. (2014). Post crisis: engage – or fly low? Brunswick Review, 8,
52-54. Retrieved from https://www.brunswickgroup.com/publications/brunswick-review/issue-8/post-crisis/.
Palen, L., Vieweg, S., Liu, S. B., & Hughes, A. L. (2009). Crisis in a Networked World: Features of Computer-Mediated Communication in the April 16, 2007, Virginia Tech Event. Social Science Computer Review, 27(4), 467-480.
doi:10.1177/0894439309332302
Procopio, C. H., & Procopio, S. T. (2007). Do You Know What It Means to Miss New Orleans? Internet Communication, Geographic Community, and Social Capital in Crisis. Journal of Applied Communication Research, 35(1), 67-87.
doi:10.1080/00909880601065722
Rindova, V. P., Williamson, I. O., Petkova, A. P., & Sever, J. M. (2005). Being Good Or Being Known: An Empirical Examination Of The Dimensions, Antecedents, And Consequences Of Organizational Reputation. Academy of Management Journal, 48(6), 1033-1049. doi:10.5465/amj.2005.19573108
Schultz, F., Utz, S., & Göritz, A. (2011). Is The Medium The Message? Perceptions Of and Reactions to Crisis Communication via Twitter, Blogs and Traditional Media. Public Relations Review, 37, 20-27. doi:10.1016/j.pubrev.2010.12.001
Smith, B. G. (2010). Socially distributing public relations: Twitter, Haiti, and interactivity in social media. Public Relations Review, 36(4), 329-335.
Sohn, Y. J., & Lariscy, R. W. (2013). Understanding Reputational Crisis: Definition, Properties, and Consequences. Journal of Public Relations Research, 26(1), 23-43. doi:10.1080/1062726x.2013.795865
Sundar, S., & Nass, C. (2001). Conceptualizing sources in online news. Journal of Communication, 51(1), 52-72. doi:10.1093/joc/51.1.52
Treem, J. W., & Leonardi, P. M. (2012). Social Media Use in Organizations: Exploring the Affordances of Visibility, Editability, Persistence, and Association. Communication Yearbook, 36, 143-189. doi:10.2139/ssrn.2129853
Tucker, L., & Melewar, T. C. (2005). Corporate Reputation and Crisis Management: The Threat and Manageability of Anti-corporatism. Corporate Reputation Review, 7(4), 377-387. doi:10.1057/palgrave.crr.1540233
Tyler, L. (1997). Liability Means Never being Able to Say Youre Sorry: Corporate Guilt, Legal Constraints, and Defensiveness in Corporate Communication. Management Communication Quarterly, 11(1), 51-73. doi:10.1177/0893318997111003
Utz, S., Schultz, F., & Glocka, S. (2013). Crisis communication online: How medium, crisis type and emotions affected public reactions in the Fukushima Daiichi nuclear disaster. Public Relations Review, 39(1), 40-46. doi:10.1016/j.pubrev.2012.09.010
Valentini, C., & Romenti, S. (2011). Blogging about crises. Journal of Communication Management, 15(4), 298-313. doi:10.1108/13632541111183398
Van der Meer, T. G. (2014). Organizational crisis-denial strategy: The effect of denial on public framing. Public Relations Review, 40(3), 537-539.
doi:10.1016/j.pubrev.2014.02.005
Van Zoonen, W., Van der Meer, T. G., & Verhoeven, J. W. (2014). Employees work-related social-media use: His masters voice. Public Relations Review, 40(5), 850-852.
doi:10.1016/j.pubrev.2014.07.001
Van Zoonen, W., & Van der Meer, T. G. (2015). The Importance of Source and Credibility Perception in Times of Crisis: Crisis Communication in a Socially Mediated Era. Journal of Public Relations Research, 27(5), 371-388.
doi:10.1080/1062726x.2015.1062382
Van Zoonen, W., Verhoeven, J. W., & Vliegenthart, R. (2016). How employees use Twitter to talk about work: A typology of work-related tweets. Computers in Human Behavior, 55, 329-339. doi:10.1016/j.chb.2015.09.021
Veil, S. R., & Ojeda, F. (2010). Establishing Media Partnerships in Crisis Response. Communication Studies, 61(4), 412-429. doi:10.1080/10510974.2010.491336 Weiner, B. (1986). An Attributional Theory of Achievement Motivation and Emotion. An
Attributional Theory of Motivation and Emotion, 159-190. doi:10.1007/978-1-4612-4948-1_6
Wendling, C., Radisch, J., & Jacobzone, S. (2013). The Use of Social Media in Risk and Crisis Communication. OECD Working Papers on Public Governance.
doi:10.1787/5k3v01fskp9s-en
Wiesenfeld, B., Raghuram, S., & Garud, R. (2001). Organizational identification among virtual workers: the role of need for affiliation and perceived work-based social
support. Journal of Management, 27(2), 213-229. doi:10.1016/s0149-2063(00)00096-9 Williams, K. D., Bourgeois, M. J., & Croyle, R. T. (1993). The effects of stealing thunder in
criminal and civil trials. Law and Human Behavior, 17(6), 597-609. doi:10.1007/bf01044684
Wright, D. K., & Hinson, M. D. (2014). An updated look at the impact of social media on public relations practice. Public Relations Journal, 8(2). Retrieved from
http://www.instituteforpr.org/wp-content/uploads/2014WrightHinson.pdf
Yates, D., & Paquette, S. (2011). Emergency knowledge management and social media technologies: A case study of the 2010 Haitian earthquake. International Journal of Information Management, 31(1), 6-13. doi:10.1016/j.ijinfomgt.2010.10.001
Zyglidopoulos, S., & Phillips, N. (1999). Responding to Reputational Crises: A Stakeholder Perspective. Corporate Reputation Review, 2(4), 333-350.
doi:10.1057/palgrave.crr.1540090
Appendix
Appendix A. Overview manipulation check of pre-test
Condition Employee Thunder Employee Stealing Thunder Organization Stealing Thunder Organization Thunder
Owner Twitter profile
Employee 6* 4* 0 1
Organization 1 1 5* 6*
Don’t know 0 1 0 0
Initial crisis reporting
